PROJECT SUMMARY The Patient Protection and Affordable Care Act of 2010 included a variety of payment and delivery system reforms to curb health spending, the most notable of which was the establishment of Medicare accountable care organizations (ACOs). Much of the initial focus of ACOs has been on enhanced primary care for beneficiaries who have multiple chronic conditions; however, there are reasons to believe that ACOs may benefit other key healthcare sectors where room for improvement exists. Specialty care is one such sector with major implications for ACOs' shared savings goals, given wide variation in its utilization and expenditures that account for nearly 50% of all health spending. While ACOs do not explicitly target specialty care, their emphasis on deeper clinical integration and greater financial stewardship could fundamentally change specialist referral patterns among primary care physicians (PCPs) and have an impact on specialists' treatment decisions when the care is of questionable value. Yet ACOs' impact on specialty care could also be limited. Most Medicare ACO contracts do not encompass specialist practices. Insofar as these providers remain weakly connected to PCPs accepting shared accountability, ACOs may have little influence on specialty care utilization and cost efficiency. They may even create perverse incentives that worsen outcomes. In this context, we propose a study to assess the impact that ACOs have on specialty care delivery. Our proposal has the following three Specific Aims. Aim 1: To measure the effects of ACOs on PCP-specialist referral networks. Using national Medicare data, we will identify patients with one of four tracer conditions (ischemic heart disease, lower extremity joint pain, and newly diagnosed breast or prostate cancer) and the provider groups that care for them. We will distinguish between groups participating in a Medicare ACO and those that are not. We will then use network analytical tools to measure specialist referral patterns in these groups before and after ACO formation. Aim 2: To assess the effects of ACOs on specialists' treatment decisions. Next, we will calculate utilization rates of high- and low-value diagnostic and therapeutic care processes across the patient cohorts identified in Aim 1. We will compare these rates among participating and non-participating provider groups before and after ACO formation. Aim 3: To determine the effects of ACOs on the efficiency of the treatment episode. Finally, we will determine Medicare payments made during treatment episodes before and after ACO formation. We will explore potential sources of savings, including component payments for physician services and hospital and ED care. Findings from our study will be directly relevant to the Agency for Healthcare Research & Quality's Research Priority Area #3, as they will inform policymakers at Medicare about the effects of ACOs on specialty care delivery.